时代的快速发展，使得数据爆炸式增长，大规模区间值属性图序列数据在许多领域广泛产生。这种数据不仅具有规模大、维数高等一般时序数据所具有的特点，还包含着两类不同性质的数据：图结构数据与顶点属性数据。这两类数据共同存在、并且相互影响，同时还以区间值的形式呈现。这些特点，使得区间值属性图序列的数据挖掘的研究更有意义、更具挑战。针对上述特点与挑战，本项目拟将基于知识指导的信息融合、基于模糊信息粒化的粒度分析、模糊逻辑与推理等理论及方法引进到大规模区间值时序属性图数据挖掘的研究中来，提出一套新的关于关联规则、聚类分析、预测分析等主要任务的挖掘方法，设计相应的高效实用的挖掘算法,以填补研究上的空白，完善大规模区间值属性图序列的数据挖掘理论，满足日益迫切的实际需要。

The rapid development of The Times makes the data explosive growth. Large-scale interval-valued attributed graph sequence data is widely produced in many fields. This kind of data not only has large scale and high dimension which are the characteristics of general temporal data, but also has graph structure data and vertex attribute data which coexist in the form of intervals and influence each other. These characteristics make the study of data mining of interval-value attributed graph sequence both more meaningful and more challenging. Aiming at the above-mentioned characteristics and challenges, this project will introduce the knowledge-guidance based information fusion, fuzzy information granulation based granular analysis, fuzzy logic and reasoning theory and methods to the study of data mining of large-scale interval-valued attributed graph sequence data, put forward a new set of methods for association rule mining, clustering analysis and forecast analysis, and design the corresponding effective practical mining algorithms. The study of this project will fill the blank of the research of large-scale interval-valued attributed graph sequence, perfect the data mining theory of large-scale interval-valued attributed graph sequence, and finally meet the increasingly urgent practical needs.